A method for decoding an image, according to the present invention, comprises the steps of: receiving image information that corresponds to a block to be decoded; performing entropy-decoding with respect to the image information that is received; deciding a transform skip mode of the block to be decoded from a plurality of transform skip mode candidates, based on the image information that is entropy-decoded; and reverse-transforming the block to be decoded based on the transform skip mode that is decided.
|
1. A method of decoding a video signal having a current block to be decoded with a decoding apparatus, comprising:
determining a scanning type of the current block, the scanning type comprising at least one of a vertical scan, a horizontal scan or a zigzag scan;
obtaining residual coefficients relating to the current block based on the determined scanning type;
obtaining inverse-quantized residual coefficients by inverse-quantizing the residual coefficients;
determining, based on a transform skip mode index specifying a transform skip mode of the current block, the transform skip mode of the current block,
wherein the transform skip mode of the current block is determined to be one of one or more transform skip mode candidates,
wherein the one or more transform skip mode candidates include at least one of a 2D transform mode, a horizontal transform mode, a vertical transform mode or a non-transform mode, and
wherein a number of the one or more transform skip mode candidates is different according to a size or a shape of the current block;
obtaining, based on the determined transform skip mode of the current block, residual samples from the inverse-quantized residual coefficients; and
decoding the current block by using the residual samples,
wherein when the transform skip mode of the current block is determined to be the non-transform mode, the residual samples are obtained by scaling the inverse-quantized residual coefficients in non-transformed rows and/or columns of the current block with a pre-determined value.
2. The method of
3. The method of
4. The method of
|
The application is a Continuation of U.S. patent application Ser. No. 14/353,288 (filed on Apr. 21, 2014), which is a National Stage Patent Application of PCT International Patent Application No. PCT/KR2012/008482 (filed on Oct. 17, 2012) under 35 U.S.C. §371, which claims priority to Korean Patent Application No. 10-2011-0106107 (filed on Oct. 17, 2011), the teachings of which are incorporated herein in their entireties by reference.
The present invention relates to image processing, and more particularly, to a transform method and a transform apparatus.
Recently, demands for high-resolution and high-quality videos, such as high-definition (HD) and ultrahigh-definition (UHD) videos, are increasing.
To provide videos with higher resolution and higher quality, the amount of video data increases. Accordingly, costs of transferring and storing video data rise so as to provide high-quality videos as compared with conventional video data processing methods. In order to solve these problems occurring with an increase in resolution and quality of video data, high-efficiency video compression techniques may be utilized.
As video data compression technology, various schemes are used such as inter prediction that is dependent on data elements of pictures other than the current picture, intra prediction that is derived from only data elements of the same decoded slice, and entropy encoding/decoding of allocating shorter codes to frequently occurring or appearing signals.
An aspect of the present invention is to provide a video encoding method and a video encoding apparatus which are capable of increasing video encoding performance.
Another aspect of the present invention is to provide a video decoding method and a video decoding apparatus which are capable of increasing video decoding performance.
Still another aspect of the present invention is to provide a transform method and a transform apparatus which are capable of increasing video encoding performance.
Yet another aspect of the present invention is to provide an inverse transform method and an inverse transform apparatus which are capable of increasing video decoding performance.
Yet another aspect of the present invention is to provide a scanning method and a scanning apparatus which are capable of increasing video encoding performance.
Yet another aspect of the present invention is to provide an inverse scanning method and an inverse scanning apparatus which are capable of increasing video decoding performance.
An embodiment of the present invention provides a video decoding method. The method may include receiving information on a picture corresponding to a decoding target block, entropy-decoding the information on the picture, determining a transform skip mode (TSM) for the decoding target block among a plurality of TSM candidates based on the entropy-decoded information on the picture, and inverse-transforming the decoding target block based on the determined TSM. Here, the TSM candidates may include at least one of a 2-directional (2D) transform mode of performing both horizontal transform and vertical transform, a horizontal transform mode of performing horizontal transform, a vertical transform mode of performing vertical transform and a non-transform mode of not performing transform.
The information on the picture may include information on a prediction mode corresponding to the decoding target block and a type of a prediction unit (PU) corresponding to the decoding target block.
When the prediction mode corresponding to the decoding target block is an inter mode and the type of the PU corresponding to the decoding target block is N×2N, N being a natural number, the vertical transform mode may be allocated a shorter codeword than the horizontal transform mode.
When the prediction mode corresponding to the decoding target block is an inter mode and the type of the PU corresponding to the decoding target block is 2N×N, N being a natural number, the TSM candidates may include the 2D transform mode, the horizontal transform mode and the non-transform mode except for the vertical transform mode.
When the prediction mode corresponding to the decoding target block is an inter mode and the type of the PU corresponding to the decoding target block is N×2N, N being a natural number, the TSM candidates may include the 2D transform mode, the vertical transform mode and the non-transform mode except for the horizontal transform mode.
When the prediction mode corresponding to the decoding target block is a short distance intra prediction (SDIP) mode and the type of the PU corresponding to the decoding target block is 2N×(1/2)N, N being a natural number that is 2 or greater, the TSM candidates may include the 2D transform mode, the horizontal transform mode and the non-transform mode except for the vertical transform mode.
When the prediction mode corresponding to the decoding target block is an SDIP mode and the type of the PU corresponding to the decoding target block is (1/2)N×2N, N being natural number that is 2 or greater, the TSM candidates may include the 2D transform mode, the vertical transform mode and the non-transform mode except for the horizontal transform mode.
The information on the picture may include information on a prediction mode corresponding to the decoding target block and a prediction direction of a PU corresponding to the decoding target block.
When the prediction mode corresponding to the decoding target block is an intra mode and the prediction direction of the PU corresponding to the decoding target block is a vertical direction, the vertical transform mode may be allocated a shorter codeword than the horizontal transform mode.
The video decoding method may further include determining a scanning mode for the decoding target block based on the determined TSM, and inverse-scanning the decoding target block based on the determined scanning mode.
The determining of the scanning mode may determine a vertical scanning mode as the scanning mode when the determined TSM is the horizontal transform mode.
The determining of the scanning mode may determine a horizontal scanning mode as the scanning mode when the determined TSM is the vertical transform mode.
Another embodiment of the present invention provides a video decoding apparatus. The apparatus may include an entropy decoding module to receive information on a picture corresponding to a decoding target block and to entropy-decode the information on the picture, and an inverse transform module to determine a TSM for the decoding target block among a plurality of TSM candidates based on the entropy-decoded information on the picture and to inverse-transform the decoding target block based on the determined TSM. Here, the TSM candidates comprise at least one of a 2D transform mode of performing both horizontal transform and vertical transform, a horizontal transform mode of performing horizontal transform, a vertical transform mode of performing vertical transform and a non-transform mode of not performing transform.
Still another embodiment of the present invention provides a video encoding method. The method may include generating a residual block corresponding to an encoding target block, determining a TSM for the encoding target block among a plurality of TSM candidates; and transforming the residual block based on the determined TSM. Here, the TSM candidates may include at least one of a 2D transform mode of performing both horizontal transform and vertical transform, a horizontal transform mode of performing horizontal transform, a vertical transform mode of performing vertical transform and a non-transform mode of not performing transform.
A prediction mode corresponding to the encoding target block may be an inter mode, and the determining of the TSM may determine the TSM based on a type of a PU corresponding to the encoding target block.
A prediction mode corresponding to the encoding target block may be an SDIP mode, and the determining of the TSM may determine the TSM based on a type of a PU corresponding to the encoding target block.
A prediction mode corresponding to the encoding target block may be an intra mode, and the determining of the TSM may determine the TSM based on intra prediction mode direction of a PU corresponding to the encoding target block.
The video encoding method may further include determining a scanning mode for the encoding target block based on the determined TSM, and scanning the encoding target block based on the determined scanning mode.
Yet another embodiment of the present invention provides a video encoding apparatus. The apparatus may include a residual block generating module to generate a residual block corresponding to an encoding target block, and a transform module to determine a TSM for the encoding target block among a plurality of TSM candidates and to transform the residual block based on the determined TSM. Here, the TSM candidates may include at least one of a 2D transform mode of performing both horizontal transform and vertical transform, a horizontal transform mode of performing horizontal transform, a vertical transform mode of performing vertical transform and a non-transform mode of not performing transform.
According to a video encoding method of the present invention, video encoding performance may be enhanced.
According to a video decoding method of the present invention, video decoding performance may be enhanced.
According to a transform/inverse transform method of the present invention, video encoding/decoding performance may be enhanced.
According to a scanning/inverse scanning method of the present invention, video encoding/decoding performance may be enhanced.
Although elements illustrated in the drawings are independently shown so as to represent different distinctive functions in a video encoding apparatus/decoding apparatus, such a configuration does not indicate that each element is constructed by a separate hardware constituent or software constituent. That is, the elements are independently arranged for convenience of description, wherein at least two elements may be combined into a single element, or a single element may be divided into a plurality of elements to perform functions. It is to be noted that embodiments in which some elements are integrated into one combined element and/or an element is divided into multiple separate elements are included in the scope of the present invention without departing from the essence of the present invention.
Hereinafter, exemplary embodiments of the invention will be described in detail with reference to the accompanying drawings. Like reference numerals in the drawings refer to like elements throughout, and redundant descriptions of like elements will be omitted herein.
The picture partitioning module 110 may divide an input picture into one or more coding units. A coding unit (CU) is a unit of encoding conducted by the video encoding apparatus and may be recursively subdivided with depth information based on a quadtree structure. A CU may have different sizes of 8×8, 16×16, 32×32 and 64×64. A CU with a maximum size is referred to as a largest coding unit (LCU), and a CU with a minimum size as a smallest coding unit (SCU).
The picture partitioning module 110 may divide a CU to generate a prediction unit (PU) and a transform unit (TU). A PU may be smaller than or the same as a CU, and may not necessarily be a square block but be a rectangular block.
Generally, intra prediction may be performed by 2N*2N or N*N blocks. Here, N is a natural number, representing a number of pixels, and 2N*2N or N*N may represent a PU size (and/or partition mode). However, in short distance intra prediction (SDIP), not only a 2N*2N PU but a subdivided PU with a size of hN*2N/2N*hN (here, h=1/2) may be also used to increase efficiency in intra prediction. When an hN*2N/2N*hN PU is used, directivity of a boundary in a block may be further reflected, and accordingly energy of a prediction error signal may be decreased to reduce bit numbers needed for encoding, thereby increasing encoding efficiency.
Inter prediction may be performed by 2N*2N, 2N*N, N*2N or N*N blocks. Here, N is a natural number, representing a number of pixels, and 2N*2N, 2N*N, N*2N or N*N may represent a PU size (and/or partition mode). Further, inter prediction may be performed by 2N×nU, 2N×nD, nL×2N or nR×2N PUs, in addition to the 2N*2N, 2N*N, N*2N or N*N PUs, in order to enhance efficiency in inter prediction. Here, 2N×nU, 2N×nD, nL×2N or nR×2N may represent a PU size (and/or partition mode). In 2N×nU and 2N×nD partition modes, a PU may have a size of 2N×(1/2)N or 2N×(3/2)N, while in nL×2N and nR×2N partition modes, a PU may have a size of (1/2)N×2N or (3/2)N×2N.
In an inter prediction mode, the inter prediction module 120 may perform motion estimation (ME) and motion compensation (MC). The inter prediction module 120 may generate a prediction block based on information on at least one of previous and subsequent pictures of the current picture.
The inter prediction module 120 may perform motion estimation based on a divided prediction target block and at least one reference block stored in the memory 155. The inter prediction module 120 may generate motion information including a motion vector (MV), a reference block index and a prediction mode as a result of motion estimation.
Further, the inter prediction module 120 may perform motion compensation using the motion information and the reference block. Here, the inter prediction module 120 may generate and output a prediction block corresponding an input block from the reference block.
In an intra prediction mode, the intra prediction module 125 may generate a prediction block based on information on a pixel in the current picture. In the intra prediction mode, the intra prediction module 125 may perform prediction for a current block based on a prediction target block and a reconstructed block previously reconstructed via transformation and quantization. Here, the reconstructed block may be a reconstructed picture that has not been subjected to the filter module 150.
In the inter prediction mode or intra prediction mode described above, prediction may be performed on a prediction target block to generate a prediction block. Here, a residual block may be generated by differentiation between the prediction target block and the generated prediction block.
The transform module 130 may transform a residual block by a TU to generate a transform coefficient. A TU may have a tree structure within maximum and minimum sizes. It may be indicated through a flag whether a current block is partitioned into sub-blocks by each TU. The transform module 130 may perform transform based on a discrete cosine transform (DCT) and/or discrete sine transform (DST).
The quantization module 135 may quantize values transformed by the transform module 130. A quantization coefficient may change based on a block or importance of a picture. The quantized transform coefficient may be provided to the rearrangement module 160 and the dequantization module 140.
The rearrangement module 160 may arrange a two-dimensional (2D) block of the quantized transform coefficients into a one-dimensional (1D) vector of transform coefficients by scanning so as to enhance efficiency in entropy encoding. The rearrangement module 160 may change scanning order based on stochastic statistics to enhance entropy encoding efficiency.
The entropy encoding module 165 may entropy-encode the values obtained by the rearrangement module 160. In entropy encoding, a more frequently occurring syntax element value may be allocated a codeword of smaller bit numbers, while a less frequently occurring syntax element value may be allocated a codeword of more bit numbers. Thus, a size of a bit string for symbols to be encoded may be reduced to enhance video encoding compression performance. Various encoding methods, such as exponential Golomb coding, context-adaptive variable length coding (CAVLC) and/or context-adaptive binary arithmetic coding (CABAC), may be used for entropy encoding. The encoded information may be formed into a compressed bitstream and be transferred or stored through a network abstraction layer (NAL).
The dequantization module 140 may dequantize the transform coefficients quantized by the quantization module 135, and the inverse transform module 145 may generate a reconstructed residual block to inverse-transform the dequantized transform coefficients. The reconstructed residual block may be merged with the prediction block generated by the inter prediction module 120 or the intra prediction module 125 to generate a reconstructed block. The reconstructed block may be provided to the intra prediction module 125 and the filter module 150.
The filter module 150 may filter the reconstructed residual block using a deblocking filter, a sample adaptive offset (SAO) and/or an adaptive loop filter (ALF). The deblocking filter may filter the reconstructed block so as to remove a distortion on boundaries between blocks occurring in encoding and decoding. The SAO is a loop filtering process to be performed on the residual block via the deblocking filter to reconstruct an offset difference from an original picture by a pixel. A band offset and an edge offset may be used as the SAO. The band offset may divide a pixel into 32 bands according to intensity and apply offsets to two divided groups of 16 bands on an edge area and 16 bands in a central area. The ALF may perform filtering so as to minimize an error between the prediction target block and the finally reconstructed block. The ALF may perform filtering based on a value obtained by comparing the reconstructed block filtered by the deblocking filter with the current prediction target block, and filter coefficient information on the ALF may be loaded onto a slice header and transferred from the encoding apparatus to the decoding apparatus.
The memory 155 may store the finally reconstructed block via the filter module 150, and the finally reconstructed block may be provided to the inter prediction module 120 performing inter prediction.
The entropy decoding module 210 may receive a compressed bitstream from an NAL. The entropy decoding module 210 may entropy-decode the received bitstream, and also entropy-decode a prediction mode and motion vector information if the bitstream includes the prediction mode and the motion vector information. When entropy decoding is used, a more frequently occurring syntax element value may be allocated a codeword of smaller bit numbers, while a less frequently occurring syntax element value may be allocated a codeword of more bit numbers. Thus, a size of a bit string for symbols to be encoded may be reduced to enhance video encoding compression performance.
An entropy-decoded transform coefficient or residual signal may be provided to the rearrangement module 215. The rearrangement module 215 may inverse-scan the decoded transform coefficient or residual signal to generate a 2D block of transform coefficients.
The dequantization module 220 may dequantize the rearranged transform coefficients. The inverse transform module 225 may inverse-transform the dequantized transform coefficients to generate a residual block.
The residual block may be merged with a prediction block generated by the inter prediction module 230 or intra prediction module 235 to generate a reconstructed block. The reconstructed block may be provided to the intra prediction module 235 and the filter module 240. The inter prediction module 230 and the intra prediction module 235 performs operations the same as or equivalent to those of the inter prediction module 120 and the intra prediction module 125 of the video encoding apparatus, and thus descriptions thereof will be omitted herein.
The filter module 240 may filter the reconstructed block using a deblocking filter, an SAO and/or an ALF. The deblocking filter may filter the reconstructed block to remove a distortion on a boundary between blocks that occurs in encoding and decoding. The SAO may be applied to the reconstructed block filtered by the deblocking filter by a pixel to reduce a difference from an original picture. The ALF may filter the reconstructed block via the SAO so as to minimize an error between the prediction target block and the finally reconstructed block.
The memory 245 may store the finally reconstructed block obtained through the filter module 240, and the stored reconstructed block may be provided to the inter prediction module 230 performing inter prediction.
Hereinafter, a block may refer to a video encoding and decoding unit. Thus, in this description, a block may mean a CU, PU, TU and the like. Also, a encoding/decoding target block may collectively include a transform/inverse transform target block, if transform/inverse transform is conducted; a prediction target block, if prediction is conducted; and the like.
As described above with reference to
The encoding apparatus and the decoding apparatus may perform 2-directional (2D) transform including vertical transform and horizontal transform. However, when vertical and horizontal signals have remarkably different characteristics, vertical transform or horizontal transform may be omitted. Also, the entire transform process may be omitted for a sparse signal. Such transform methods may reduce complexity in the decoding apparatus and improve encoding efficiency.
Hereinafter, a transform mode involving both horizontal transform and vertical transform is referred to as a “2D transform mode.” A transform mode involving horizontal transform only without vertical transform is referred to as a “horizontal transform mode,” and a transform mode involving vertical transform only without horizontal transform is referred to as a “vertical transform mode.” Further, a transform mode involving neither horizontal transform nor vertical transform is referred to as a “non-transform mode.” Here, the non-transform mode may be also referred to as a “transform bypass mode.”
Square blocks 310 to 340 shown in
A transform target block 310 may be subjected to both vertical transform and horizontal transform. Thus, a transform mode for the block 310 may correspond to the 2D transform mode. A transform target block 320 may be subjected to horizontal transform only without vertical transform. Thus, a transform mode for the block 320 may correspond to the horizontal transform mode. In this case, since transform is performed on rows, not on columns, a transform method in the horizontal transform mode may be also referred to as “transform on rows only.” A transform target block 330 may be subjected to vertical transform only without horizontal transform. Thus, a transform mode for the block 330 corresponds to the vertical transform mode. In this case, since transform is performed on columns, not on rows, a transform method in the vertical transform mode may be also referred to as “transform on columns only.” A transform target block 340 may not be subjected to transform. Thus, a transform mode for the block 340 is the non-transform mode.
In the foregoing transform modes, vertical transform and/or horizontal transform may be or may not be omitted. Thus, these transform modes may be also referred to as a transform skip mode (TSM). That is, the transform skip mode may include the 2D transform mode, the horizontal transform mode, the vertical transform mode and the non-transform mode. The 2D transform mode, the horizontal transform mode, the vertical transform mode and/or the non-transform mode may be used as candidates for the transform skip mode for a transform target block.
In one exemplary embodiment, at least one of the 2D transform mode, the horizontal transform mode, the vertical transform mode and the non-transform mode may be used as a transform skip mode candidate for a transform target block. Here, one transform skip mode selected from a plurality of transform skip mode candidates may be applied to one transform target block. The encoding apparatus may select a transform skip mode having a smallest cost value in view of rate-distortion optimization (RDO) among the transform skip mode candidates. Here, the encoding apparatus may transform the transform target block based on the selected transform skip mode. That is, the encoding apparatus may apply one selected transform skip mode of the 2D transform mode, the horizontal transform mode, the vertical transform mode and/or the non-transform mode to the transform target block.
In addition, the encoding apparatus may encode information on the selected transform skip mode and transmit the information to the decoding apparatus. The transform skip mode may be determined by a CU or TU. Here, when the transform skip mode is determined by a CU, the information may be transmitted by a CU. When the transform skip mode is determined by a TU, the information may be transmitted by a TU.
For instance, the information on the transform skip mode may be transmitted to the decoding apparatus through a transform skip mode index. The transform skip mode index may be an index indicating the transform skip mode to be applied to the transform target block among the transform skip mode candidates. The transform skip mode index may be allocated an index value based on the transform skip mode. Here, the 2D transform mode, the horizontal transform mode and the vertical transform mode may correspond different index values.
The decoding apparatus may decode the information about the transform skip mode (for example, the encoded transform skip mode index) which is received from encoding apparatus. Here, the decoding apparatus may derive the transform skip mode to be applied to the transform target block based on the decoded information. The decoding apparatus may transform the transform target block according to the derived transform skip mode. That is, the decoding apparatus may apply one derived transform skip mode of the 2D transform mode, the horizontal transform mode, the vertical transform mode and/or the non-transform mode to the transform target block.
Referring to
Referring back to
Further, the encoding apparatus may encode information on the transform skip mode applied to the transform target block and transmit the information to the decoding apparatus. For example, the information may be transmitted to the decoding apparatus through a transform skip mode index. Here, as described above, considering probabilities of transform skip modes, the encoding apparatus may allocate a short codeword to a more likely transform skip mode and a long codeword to a less likely transform skip mode. A method of allocating a codeword for a transform skip mode according to an exemplary embodiment will be described in detail.
The decoding apparatus may decode a bitstream including the information about the transform skip mode (for example, the encoded transform skip mode index) which is received from encoding apparatus. In the bitstream received from the encoding apparatus, a short codeword may be allocated to a more likely transform skip mode, and a long codeword may be allocated to a less likely transform skip mode. A method of allocating a codeword for a transform skip mode according to an exemplary embodiment will be described in detail.
Referring to
Referring back to
Meanwhile, in the embodiments illustrated in
TABLE 1
Row
Column
Codeword
trans-
trans-
(CABAC and/
TSM
formation
formation
or CAVLC)
Note
TS0
O
O
1
2D transform
TS1
O
—
01
1D transform
TS2
—
O
001
1D transform
TS3
—
—
000
Non-transform
In Table 1, TS0 represents the 2D transform mode. TS1 represents the horizontal transform mode, and TS2 represents the vertical transform mode. TS3 represents the non-transform mode. Here, both the horizontal transform mode and the vertical transform mode may correspond to a 1D transform mode.
For example, referring to Table 1, if the 2D transform mode most frequently happens, the 2D transform mode may be allocated a codeword “1.” Likewise, according to frequency, the horizontal transform mode may be allocated a codeword “01,” the vertical transform mode a codeword “001,” and the non-transform mode a codeword “000.”
Even when vertical transform and/or horizontal transform is omitted depending on transform skip modes, the same quantization matrix may be applied as in the 2D transform mode. In this case, the encoding apparatus and the decoding apparatus may perform scaling on values in rows and/or columns not subjected to transform, which may be represented by Equation 1.
y=(x*scale+offset)>>shift [Equation 1]
Here, x may be an element in a non-transformed row and/or column, and y may be a scaled value. “scale” may be a scaling factor. “offset” may be an offset value applied in scaling, and “shift” may be a bit shifting value applied in scaling. Here, “offset” and “shift” may have the same values as an offset value and a bit transfer value applied when transform is not omitted, for example, in the 2D transform mode.
Further, in Equation 1, the scaling factor applied to the encoding apparatus and the decoding apparatus may be determined depending on a TU size. In one exemplary embodiment, the scaling factor according to the TU size may be set as listed in Table 2.
TABLE 2
N
4
8
16
32
Scale
128
181
256
362
Here, N (and/or N×N) may be a TU size, and scale may be a scaling factor. Referring to
As mentioned above, a PU may not necessarily have a square shape but have a rectangular shape. For example, in the inter mode, a PU may have a 2N*N, N*2N, 2N×nU, 2N×nD, nL×2N or nR×2N size (and/or shape). In SDIP, a PU may have a 2N*(1/2)N or (1/2)N*2N size (and/or shape). In this instance, since a particular transform skip mode may be less likely to happen, the encoding apparatus and the decoding apparatus may not use the less likely transform skip mode as a transform skip mode candidate, thereby enhancing encoding/decoding performance. Alternatively, the encoding apparatus may allocate a short codeword to the less likely transform skip mode, thereby enhancing encoding/decoding performance. Accordingly, a method of determining a transform skip mode candidate and a method of allocating a codeword for a transform skip mode according to a PU size (and/or form) may be provided.
Meanwhile, in the inter mode, probabilities of the horizontal transform mode and the vertical transform mode may vary on PU forms. Thus, different codewords may be allocated to transform skip modes (and/or transform skip mode indexes) depending on PU forms. That is, codewords allocated to transform skip modes (and/or transform skip mode indexes) may be determined based on PU forms.
In one exemplary embodiment, when a PU has a form of N*2N 623, energy compaction effect of horizontal transform may be smaller than energy compaction effect of vertical transform. Thus, the vertical transform mode may have a higher probability than the horizontal transform mode. In Table 1, the horizontal transform mode is allocated the codeword “01” and the vertical transform mode is allocated the codeword “001,” that is, a more likely transform skip mode is allocated a longer codeword. Thus, in the PU with the form of N*2N 623, the codeword for the horizontal transform mode and the codeword for the vertical transform mode are reset, thereby enhancing encoding performance. Table 3 illustrates a method of allocating codewords to transform skip modes in the PU with the form of N*2N 623 according to an exemplary embodiment.
TABLE 3
Row
Column
Codeword
trans-
trans-
(CABAC and/
TSM
formation
formation
or CAVLC)
Note
TS0
O
O
1
2D transform
TS1
O
—
001
1D transform
TS2
—
O
01
1D transform
TS3
—
—
000
Non-transform
In Table 3, TS0 represents the 2D transform mode. TS1 represents the horizontal transform mode, and TS2 represents the vertical transform mode. TS3 represents the non-transform mode. Here, both the horizontal transform mode and the vertical transform mode may correspond to a 1D transform mode.
Referring to Table 3, the horizontal transform mode may be allocated a codeword “001,” and the vertical transform mode may be allocated a codeword “01.” As described above, in the PU with the form of N*2N 623, the vertical transform mode may have a higher probability than the horizontal transform mode, and thus the vertical transform mode may be allocated a shorter code than the horizontal transform mode.
Although Table 3 is described based on the PU with the form of N*2N 623, the present invention is not limited thereto. For example, in a PU with a form of nL×2N 627 or nR×2N 628 in addition to N*2N 623, the vertical transform mode may also have a higher probability than the horizontal transform mode. Accordingly, the vertical transform mode may be allocated a shorter code than the horizontal transform mode.
On the other hand, in PUs with 2N*N 622, 2N×nU 625 and 2N×nD 626 forms, the horizontal transform mode may have a higher probability than the vertical transform mode. Accordingly, the horizontal transform mode may be allocated a shorter code than the vertical transform mode. For example, in the PU with the 2N*N 622, 2N×nU 625 and 2N×nD 626 forms, the same codeword allocation method as in Table 1 may be used.
Meanwhile, as described above, since probabilities of the horizontal transform mode and the vertical transform mode in the inter mode may vary on PU forms, a number of transform skip mode candidates may be determined differently based on PU forms. That is, transform skip mode candidates for a transform target block may be determined based on a PU form corresponding to the transform target block.
In one exemplary embodiment, when a PU has a 2N*N 622 form, energy compaction effect of vertical transform may be smaller than energy compaction effect of horizontal transform, and thus the vertical transform mode may have a lower probability than the horizontal transform mode. Thus, in the PU with the 2N*N 622 form, the 2D transform mode, the horizontal transform mode and the non-transform mode may be used as transform skip mode candidates for a transform target block, excluding the vertical transform mode. In this case, one transfer skip mode among the 2D transform mode, the horizontal transform mode and the non-transform mode may be applied to the transform target block. Table 4 illustrates a method of allocating codewords to transform skip modes when the 2D transform mode, the horizontal transform mode and the non-transform mode are used as transform skip mode candidates according to an exemplary embodiment.
TABLE 4
Row
Column
Codeword
trans-
trans-
(CABAC and/
TSM
formation
formation
or CAVLC)
Note
TS0
O
O
0
2D transform
TS1
O
—
10
1D transform
TS3
—
—
11
Non-transform
In Table 4, TS0 represents the 2D transform mode, TS1 represents the horizontal transform mode, and TS3 represents the non-transform mode. Here, the horizontal transform mode may correspond to a 1D transform mode. Referring to Table 4, in the PU with the 2N*N 622 form, the 2D transform mode, the horizontal transform mode and the non-transform mode may be used as transform skip mode candidates.
Although Table 4 is described based on the PU with the form of 2N*N 622, the present invention is not limited thereto. For example, in PUs with forms of 2N×nU 625 and 2N×nD 626 in addition to 2N*N 622, the vertical transform mode may also have a lower probability than the horizontal transform mode. Accordingly, the 2D transform mode, the horizontal transform mode and the non-transform mode may be used as transform skip mode candidates for a transform target block, excluding the vertical transform mode.
Alternatively, in the PU with the form of N*2N 623, since energy compaction effect of horizontal transform may be smaller than energy compaction effect of vertical transform, the horizontal transform mode may have a lower probability than the vertical transform mode. Thus, in the PU with the form of N*2N 623, the 2D transform mode, the vertical transform mode and the non-transform mode may be used as transform skip mode candidates for a transform target block, excluding the horizontal transform mode. In this case, one transform skip mode among the 2D transform mode, the vertical transform mode and the non-transform mode may be applied to the transform target block. Table 5 illustrates a method of allocating codewords to transform skip modes when the 2D transform mode, the vertical transform mode and the non-transform mode are used as transform skip mode candidates according to an exemplary embodiment.
TABLE 5
Row
Column
Codeword
trans-
trans-
(CABAC and/
TSM
formation
formation
or CAVLC)
Note
TS0
O
O
0
2D transform
TS2
—
O
10
1D transform
TS3
—
—
11
Non-transform
In Table 5, TS0 represents the 2D transform mode, TS2 represents the vertical transform mode, and TS3 represents the non-transform mode. Here, the vertical transform mode may correspond to a 1D transform mode. Referring to Table 5, in the PU with the form of N*2N 623, the 2D transform mode, the vertical transform mode and the non-transform mode may be used as transform skip mode candidates.
Although Table 5 is described based on the PU with the form of N*2N 623, the present invention is not limited thereto. For example, in a PU with a form of nL×2N 627 or nR×2N 628 in addition to N*2N 623, the horizontal transform mode may also have a lower probability than the vertical transform mode. Accordingly, the 2D transform mode, the vertical transform mode and the non-transform mode may be also used as transform skip mode candidates for a transform target block, excluding the horizontal transform mode.
In the foregoing embodiments illustrated in Tables 3 to 5, bit numbers used for encoding transform skip modes (and/or transform skip mode indexes) may be reduced. According to, encoding/decoding performance may be enhanced.
In SDIP, since probabilities of the horizontal transform mode and the vertical transform mode may vary on PU forms, a number of transform skip mode candidates may be determined differently based on PU forms. That is, transform skip mode candidates for a transform target block may be determined based on a PU form corresponding to the transform target block.
In one exemplary embodiment, when a PU has a 2N*(1/2)N 727 form, energy compaction effect of vertical transform may be smaller than energy compaction effect of horizontal transform, and thus the vertical transform mode may have a lower probability than the horizontal transform mode. Thus, in the PU with the 2N*(1/2)N 727 form, the 2D transform mode, the horizontal transform mode and the non-transform mode may be used as transform skip mode candidates for a transform target block, excluding the vertical transform mode. In this case, one transfer skip mode among the 2D transform mode, the horizontal transform mode and the non-transform mode may be applied to the transform target block. A method of allocating codewords to transform skip modes when the 2D transform mode, the horizontal transform mode and the non-transform mode are used as transform skip mode candidates has been described above in Table 4, and thus a description thereof will be omitted herein.
Alternatively, in a PU with a form of (1/2)N*2N 725, since energy compaction effect of horizontal transform may be smaller than energy compaction effect of vertical transform, the horizontal transform mode may have a lower probability than the vertical transform mode. Thus, in the PU with the form of (1/2)N*2N 725, the 2D transform mode, the vertical transform mode and the non-transform mode may be used as transform skip mode candidates for a transform target block, excluding the horizontal transform mode. In this case, one transform skip mode among the 2D transform mode, the vertical transform mode and the non-transform mode may be applied to the transform target block. A method of allocating codewords to transform skip modes when the 2D transform mode, the vertical transform mode and the non-transform mode are used as transform skip mode candidates has been described above in Table 5, and thus a description thereof will be omitted herein.
In the foregoing embodiments, bit numbers used for encoding transform skip modes (and/or transform skip mode indexes) may be reduced. According to, encoding/decoding performance may be enhanced.
As described above with reference to
Referring to
Probabilities of the horizontal transform mode and the vertical transform mode may vary on an intra prediction mode (and/or prediction direction) of a PU corresponding to a transform target block. Thus, a different codeword may be allocated to a transform skip mode (and/or transform skip mode index) based on the intra prediction mode (and/or prediction direction) of the PU. That is, a codeword allocated to a transform skip mode (and/or transform skip mode index) may be determined based on the intra prediction mode (and/or prediction direction) of the PU corresponding to the transform target block.
In one exemplary embodiment, when the intra prediction mode of the PU is the vertical mode, energy compaction effect of horizontal transform may be smaller than energy compaction effect of vertical transform. Thus, in this case, the vertical transform mode may have a higher probability than the horizontal transform mode. In the embodiment illustrated with reference to Table 1, the horizontal transform mode is allocated the codeword “01” and the vertical transform mode is allocated the codeword “001,” that is, a more likely transform skip mode is allocated a longer codeword. Thus, when the intra prediction mode of the PU is the vertical mode, the codeword for the horizontal transform mode and the codeword for the vertical transform mode are reset, thereby enhancing encoding performance. That is, when the intra prediction mode of the PU is the vertical mode, the vertical transform mode may have a higher probability than the horizontal transform mode, the vertical transform mode may be allocated a shorter code than the horizontal transform mode. An embodiment of allocating a shorter codeword to the vertical transform mode than to the horizontal transform mode is similar to the embodiment illustrated in Table 3, and thus a description thereof is omitted herein.
Alternatively, when the intra prediction mode of the PU corresponding to the transform target block is the horizontal mode, the horizontal transform mode may have a higher probability than the vertical transform mode. Thus, in this case, the horizontal transform mode may be allocated a shorter code than the vertical transform mode. For example, when the intra prediction mode of the PU corresponding to the transform target block is the horizontal mode, the same codeword allocation method as in Table 1 may be used.
In the embodiment of
As described above in
Here, the encoding apparatus and the decoding apparatus may determine a scanning method (and/or scanning order) based on a transform skip mode. That is, according to exemplary embodiments of the present invention, different scanning methods (and/or scanning orders) may be used based on a transform skip mode for a transform target block.
In one exemplary embodiment, when the transform skip mode is the horizontal transform mode, residual signals are more likely to remain in the vertical direction. Thus, when the transform skip mode for the transform target block is the horizontal transform mode, vertical scanning 920 may be used for the transform target block. When the transform skip mode is the vertical transform mode, residual signal are more likely to remain in the horizontal direction. Thus, when the transform skip mode for the transform target block is the vertical transform mode, horizontal scanning 910 may be used for the transform target block. In transform skip modes other than the horizontal transform mode and the vertical transform mode, zigzag scanning 930 may be used to perform scanning.
Referring to
In
C(n,n)=T(n,n)×B(n,n)×T(n,n)T [Equation 2]
Here, C(n,n) is an n*n transform coefficient matrix, T(n,n) is an n*n transformation kernel matrix, and B(n,n) is an n*n matrix of a residual block.
When a transform coefficient is generated via transformation, the encoding apparatus may quantize the generated transform coefficient.
It may be determined through RDO which is transmitted among the residual block and the transform coefficient. When prediction is properly done, the residual block, that is, the residual signal, may be transmitted without transcoding. The encoding apparatus may compare cost functions before/after transcoding and select a method involving minimum costs. Here, the encoding apparatus may transmit information on a type of a signal (residual signal or transform coefficient) transmitted with respect to the current block to the decoding apparatus.
Transform processes have been illustrated in the foregoing embodiments, and thus descriptions thereof are omitted herein.
Referring back to
When scanning is performed, the encoding apparatus may entropy-encode the scanned transform coefficient and side information (for example, information on an inter prediction mode of the current block) (S1040). The encoded information may be formed into a compressed bitstream and be transferred or stored through an NAL.
Although the encoding method is described with a series of stages based on the flowchart in
Referring to
In
As described above, in inverse scanning, the decoding apparatus may determine a scanning method (and/or scanning order) based on a transform skip mode. A method of determining a scanning order based on a transform skip mode has been described above, and thus a description thereof is omitted herein.
Referring back to
B(n,n)=T(n,n)×C(n,n)×T(n,n)T [Equation 3]
Inverse transformation has been described above, and thus a description thereof is omitted herein.
When the residual block is generated, the decoding apparatus may generate a reconstructed block based on the generated residual block (S1140). As described above, the decoding apparatus may perform inter prediction and/or intra prediction on a decoding target block to generate a prediction block corresponding to the decoding target block. Here, the decoding apparatus may merge a pixel value of the prediction block and a pixel value of the residual block by a pixel, thereby generating the reconstructed block.
Although the decoding method is described with a series of stages based on the flowchart in
Although methods have been described with a series of stages or blocks based on the flowcharts in the aforementioned embodiments, the present invention is not limited to the foregoing sequence of the stages. Some stages may be carried out in different order from described above or parallel at the same time. Also, it will be understood by those skilled in the art that the stages illustrated in the flowcharts are not exclusive, additional stages may be included in the flowchart, or one or more stages may be deleted from the flowcharts without affecting the scope of the present invention.
The present invention has been described with reference to the exemplary embodiments, and the foregoing embodiments include various aspects of examples. Although all possible combinations may not be mentioned to illustrate various aspects, it will be appreciated by those skilled in the art that changes, modifications and alternatives may be made in these exemplary embodiments without departing from the principles and spirit of be the invention, the scope of which is defined in the appended claims and their equivalents.
Kim, Joo Young, Lee, Bae Keun, Kwon, Jae Cheol
Patent | Priority | Assignee | Title |
Patent | Priority | Assignee | Title |
7092576, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Bitplane coding for macroblock field/frame coding type information |
7099515, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Bitplane coding and decoding for AC prediction status information |
7162093, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Slice-layer in video codec |
7317839, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Chroma motion vector derivation for interlaced forward-predicted fields |
7352905, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Chroma motion vector derivation |
7369709, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Conditional lapped transform |
7412102, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Interlace frame lapped transform |
7469011, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Escape mode code resizing for fields and slices |
7529302, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Four motion vector coding and decoding in bi-directionally predicted interlaced pictures |
7567617, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Predicting motion vectors for fields of forward-predicted interlaced video frames |
7577198, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Number of reference fields for an interlaced forward-predicted field |
7577200, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Extended range variable length coding/decoding of differential motion vector information |
7590179, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Bitplane coding of prediction mode information in bi-directionally predicted interlaced pictures |
7599438, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Motion vector block pattern coding and decoding |
7606308, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Signaling macroblock mode information for macroblocks of interlaced forward-predicted fields |
7606311, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Macroblock information signaling for interlaced frames |
7609762, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Signaling for entry point frames with predicted first field |
7616692, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Hybrid motion vector prediction for interlaced forward-predicted fields |
7620106, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Joint coding and decoding of a reference field selection and differential motion vector information |
7623574, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Selecting between dominant and non-dominant motion vector predictor polarities |
7630438, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Direct mode motion vectors for Bi-directionally predicted interlaced pictures |
7664177, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Intra-coded fields for bi-directional frames |
7680185, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Self-referencing bi-directionally predicted frames |
7688894, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Scan patterns for interlaced video content |
7724827, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Multi-layer run level encoding and decoding |
7782954, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Scan patterns for progressive video content |
7822123, | Oct 06 2004 | Microsoft Technology Licensing, LLC | Efficient repeat padding for hybrid video sequence with arbitrary video resolution |
7839930, | Nov 13 2003 | Microsoft Technology Licensing, LLC | Signaling valid entry points in a video stream |
7852919, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Field start code for entry point frames with predicted first field |
7852936, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Motion vector prediction in bi-directionally predicted interlaced field-coded pictures |
7924920, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Motion vector coding and decoding in interlaced frame coded pictures |
7924921, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Signaling coding and display options in entry point headers |
7961786, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Signaling field type information |
8005137, | Mar 25 2005 | Samsung Electronics Co., Ltd. | Video coding and decoding method using weighted prediction and apparatus for the same |
8009739, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Intensity estimation/compensation for interlaced forward-predicted fields |
8014450, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Flexible range reduction |
8064520, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Advanced bi-directional predictive coding of interlaced video |
8085844, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Signaling reference frame distances |
8107531, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Signaling and repeat padding for skip frames |
8116380, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Signaling for field ordering and field/frame display repetition |
8213779, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Trick mode elementary stream and receiver system |
8345754, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Signaling buffer fullness |
8396123, | Mar 25 2005 | Samsung Electronics Co., Ltd. | Video coding and decoding method using weighted prediction and apparatus for the same |
8488668, | Jun 15 2007 | Qualcomm Incorporated | Adaptive coefficient scanning for video coding |
8582659, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Determining a decoding time stamp from buffer fullness |
8625669, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Predicting motion vectors for fields of forward-predicted interlaced video frames |
8625670, | Mar 17 2008 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding image |
8625680, | Sep 07 2003 | Microsoft Technology Licensing, LLC | Bitstream-controlled post-processing filtering |
8687709, | Sep 07 2003 | Microsoft Technology Licensing, LLC | In-loop deblocking for interlaced video |
8711950, | Feb 14 2008 | Sony Corporation | Apparatus and method for adapted deblocking filtering strength |
8885714, | Jan 14 2010 | Texas Instruments Incorporated | Method and system for intracoding in video encoding |
20030016755, | |||
20050052294, | |||
20050053146, | |||
20060018387, | |||
20060104521, | |||
20070133891, | |||
20070286283, | |||
20080137738, | |||
20090097571, | |||
20090207911, | |||
20090232217, | |||
20090238271, | |||
20100054334, | |||
20100195715, | |||
20110058608, | |||
20110090952, | |||
20110090954, | |||
20110090967, | |||
20110268183, | |||
20120281928, | |||
CN101040533, | |||
CN101895756, | |||
CN102025986, | |||
CN1741609, | |||
EP2008035842, | |||
EP2086239, | |||
EP2154898, | |||
JP2007516640, | |||
JP2009194617, | |||
KP1020090129939, | |||
KR1020090099234, | |||
RU2341035, |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
Mar 19 2014 | LEE, BAE KEUN | KT Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 040637 | /0410 | |
Mar 19 2014 | KWON, JAE CHEOL | KT Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 040637 | /0410 | |
Mar 19 2014 | KIM, JOO YOUNG | KT Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 040637 | /0410 | |
Dec 15 2016 | KT Corporation | (assignment on the face of the patent) | / |
Date | Maintenance Fee Events |
Sep 25 2020 | M1551: Payment of Maintenance Fee, 4th Year, Large Entity. |
Date | Maintenance Schedule |
May 23 2020 | 4 years fee payment window open |
Nov 23 2020 | 6 months grace period start (w surcharge) |
May 23 2021 | patent expiry (for year 4) |
May 23 2023 | 2 years to revive unintentionally abandoned end. (for year 4) |
May 23 2024 | 8 years fee payment window open |
Nov 23 2024 | 6 months grace period start (w surcharge) |
May 23 2025 | patent expiry (for year 8) |
May 23 2027 | 2 years to revive unintentionally abandoned end. (for year 8) |
May 23 2028 | 12 years fee payment window open |
Nov 23 2028 | 6 months grace period start (w surcharge) |
May 23 2029 | patent expiry (for year 12) |
May 23 2031 | 2 years to revive unintentionally abandoned end. (for year 12) |